Patel et al ran 9,999 computer simulations using various vancomycin doses and MICs. When the MIC was 2.0 or greater, 2 g IV q12 hours failed to achieve target drug levels 57% of the time, with 35% predicted nephrotoxicity. But...it's also true that multiple vancomycin models have been published, which are each internally valid using their own study data, but yielding widely inconsistent results when given the same inputs. Clin Infect Dis 2011;52:969-974.